Diversification of an Endemic Southeast Asian : Phylogenetic Relationships of the (Nectariniidae: Arachnothera) Author(s) :Robert G. Moyle, Sabrina S. Taylor, Carl H. Oliveros, Haw Chuan Lim, Cheryl L. Haines, Mustafa A. Rahman, and Frederick H. Sheldon Source: The Auk, 128(4):777-788. 2011. Published By: The American Ornithologists' Union URL: http://www.bioone.org/doi/full/10.1525/auk.2011.11019

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. The Auk 128(4):777−788, 2011 © The American Ornithologists’ Union, 2011. Printed in USA.

DIVERSIFICATION OF AN ENDEMIC SOUTHEAST ASIAN GENUS: PHYLOGENETIC RELATIONSHIPS OF THE SPIDERHUNTERS (NECTARINIIDAE: ARACHNOTHERA)

ROBERT G. MOYLE,1,5 SABRINA S. TAYLOR,2 CARL H. OLIVEROS,1 HAW CHUAN LIM,3 CHERYL L. HAINES,3 MUSTAFA A. RAHMAN,4 AND FREDERICK H. SHELDON3

1Biodiversity Institute and Department of Ecology and Evolutionary Biology, University of Kansas, Lawrence, Kansas 66045, USA; 2School of Renewable Natural Resources, Louisiana State University Agricultural Center, Baton Rouge, Louisiana 70803, USA; 3Museum of Natural Science and Department of Biological Sciences, Louisiana State University, Baton Rouge, Louisiana 70803, USA; and 4Faculty of Resource Science and Technology, Universiti Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia

Abstract.—The phylogeny of spiderhunters (Nectariniidae: Arachnothera) was reconstructed by comparing mitochondrial and nuclear DNA sequences of all currently recognized species and with broad geographic sampling of two particularly variable species complexes, the Little (Arachnothera longirostra) and the streaky spiderhunters (A. modesta and A. affinis). It appears to be a relatively old group, whose diversification was not caused by recent sea-level changes. However, the modern, highly sympatric distribution of the large species in the Sunda lowlands was probably a result of dispersal via recent land bridges. Within the highly variable A. longirostra group, there are substantially diverged taxa in the that should be considered different species. Within the A. affinis–modesta complex, there are three distinct species and a closely related fourth, which describe a clear allopatric distribution: A. affinis in Java, A. modesta in the rest of the Sunda lowlands (except Sabah), A. magna in the Malayan highlands and mainland Southeast , and A. everetti in the Bornean highlands and Sabah. Depending on whether mitochondrial or nuclear genes were compared, monophyly of the genus was disrupted by a single outgroup (Hypogramma hypogrammicum) or by all outgroup included in the study. The discrepancy between nuclear and mitochondrial results is probably a case of deep coalescence and will require additional markers for resolution. Received  January , accepted  August .

Key words: Arachnothera, Nectariniidae, phylogeny, Southeast Asia, spiderhunters.

Diversification d’un genre endémique de l’Asie du Sud-Est : les relations phylogénétiques d’Arachnothères (Nectariniidae : Arachnothera)

Résumé.—La phylogénie des Arachnothères (Nectariniidae : Arachnothera) a été construite en comparant les séquences d’ADN mitochondrial et nucléaire de toutes les espèces actuellement reconnues et avec un large échantillonnage géographique de deux complexes d’espèces particulièrement variables, soit Arachnothera longirostra ainsi qu’A. affinis–modesta. Il semble s’agir d’un groupe relativement ancien, dont la diversification n’a pas été causée par des changements récents du niveau de la mer. Cependant, la répartition moderne et hautement sympatrique de la grosse espèce dans les basses terres de Sunda était probablement le résultat de la dispersion via des ponts terrestres récents. Au sein du groupe hautement variable d’A. longirostra, il existe des taxons substantiellement différenciés aux Philippines qui devraient être considérés comme des espèces différentes. Pour le complexe A. affinis–modesta, il existe trois espèces distinctes et une quatrième étroitement apparentée, qui décrivent une répartition allopatrique claire : A. affinis à Java, A. modesta dans le reste des terres basses de Sunda (excepté Sabah), A. magna dans les hautes terres de Malaya et la partie continentale de l’Asie du Sud-Est, ainsi qu’A. everetti dans les hautes terres de Bornéo et Sabah. Selon que les gènes mitochondriaux ou nucléaires étaient comparés, la monophylie du genre était perturbée par un seul hors-groupe d’oiseaux (Hypogramma hypogrammicum) ou par tous les hors-groupes d’oiseaux inclus dans l’étude. L’écart entre les résultats nucléaires et mitochondriaux constitue probablement un cas de coalescence profonde qui requiert des marqueurs supplémentaires pour être résolu.

The two major island groups of Southeast Asia, the Greater are largely oceanic islands (Hall and Holloway ). With the ex- Sundas and the Philippines, have distinct biogeographic histories. ception of Palawan, which lies on the Sunda shelf, the Philippines The Greater Sundas are continental islands, but the Philippines were never connected to any other land mass and, thus, have been

5E-mail: [email protected]

The Auk, Vol. , Number , pages −. ISSN -, electronic ISSN -. ©  by The American Ornithologists’ Union. All rights reserved. Please direct all requests for permission to photocopy or reproduce article content through the University of California Press’s Rights and Permissions website, http://www.ucpressjournals. com/reprintInfo.asp. DOI: ./auk.. — 777 — 778 — MOYLE ET AL. — AUK, VOL. 128

considerably more isolated than the Greater Sundas (Heaney ). Borneo, Sumatra, Java, their associated small islands, and the Ma- lay Peninsula lie on the Sunda continental shelf. Much of the shelf is covered by a shallow sea that was exposed during periods of gla- ciation and low sea level in the Pleistocene, and the resulting large land mass known as Sundaland encompassed the islands and main- land of Southeast Asia (Whitmore ). Concomitant with low sea levels, colder climates and reduced oceanic effects caused montane in the region to descend to lower elevations (Flenley , Cannon et al. ), thus forming barriers to lowland populations and corridors among previously isolated higher montane popula- tions. The colder climate and drier conditions also caused savanna and seasonal forests to replace rainforest in interior lowland areas of Sundaland, thus further separating rainforest taxa into allopatric populations (Heaney , Flenley , Meijaard , et al. , Cannon et al. ). Recent molecular phylogenetic studies have begun to define distinct patterns of diversification within the Greater Sundas and Philippines that bear on potential connections and disconnections among populations (Gorog et al. , Campbell et al. , Jansa et al. , Ryan and Esa , Esselstyn and Oliveros , Es- selstyn et al. ). For , most phylogenetic work in the region has focused on Borneo, which is a key link between the Philippines and the Greater Sundas (Diamond and Gilpin ). Moyle et al. () first documented a major genetic discontinuity within low- land Borneo that made populations of the White-crowned Forktail FIG. 1. Map of spiderhunter (Arachnothera) distributions in Southeast (Enicurus leschenaulti), a saxicoline thrush (Sangster et al. ), Asia from Cheke and Mann (2008) and Kennedy et al. (2000). The west- in Sarawak (northwest Borneo) more closely related to popula- ern extensions of Arachnothera longirostra in and A. magna in the tions in Sumatra and the Malay Peninsula than to populations in eastern Himalayas are not shown. The distribution of A. modesta extends Sabah (northeast Borneo). Subsequent work on additional lowland farther north than shown, into far southwest and . The bird species has expanded the list of taxa that exhibit this pattern inset provides a more precise estimate of the distribution of A. affinis and (Sheldon et al. b; Lim et al. , ). Studies of widespread A. modesta on Borneo (Sheldon et al. 2001). Southeast Asian avian taxa have also found that Philippine popula- tions tend to be more distinct genetically from their Sundaic rela- Two of the spiderhunter species complexes comprise con- tives than previously thought. As such, they merit consideration as siderable geographic variability and taxonomic uncertainty. The endemic Philippine species rather than simply polymorphic exten- (A. longirostra) is the most widespread spe- sions of continental Southeast Asian species (Sheldon et al. b; cies in the genus, ranging from India to the Philippines. Molec- Lim et al. , ; Lohman et al. ; Oliveros and Moyle ). ular phylogenetic comparisons have shown a deep genetic split The genusArachnothera , the spiderhunters, presents an excel- between the population on Mindanao in the southern Philippines lent opportunity to examine these patterns further and investigate and populations in Sundaland and mainland Asia (Lohman et al. the processes responsible for avian diversification in Sundaland and , Rahman et al. ). Rahman et al. () also found that the Philippines. Spiderhunters are members of the sunbird family the population on Palawan is distinct, even though Palawan is a (Nectariniidae) and include widespread as well as island endemic Sundaic island adjacent to Borneo. Both of these studies included species across southern Asia, the Greater Sunda Islands, and portions only mitochondrial DNA data and samples from single sites in the of the Philippines (Fig. ). They reach their greatest diversity on Bor- Philippines (Mindanao and Palawan). Here, we compare a larger neo. Of the  currently recognized species (Dickinson ), seven number of A. longirostra populations using additional molecular are confined to Sundaland, one has a Sundaic as well as an Asian markers to provide a broader biogeographic assessment of the spe- mainland population (A. magna), one ranges from mainland South- cies complex. east Asia across Sundaland and into the oceanic Philippines (A. lon- The complex consisting of A. affinis and A. modesta (here- girostra), and one is confined to the oceanic Philippines (A. clarae). after “A. affinis–modesta complex”) comprises five subspecies Morphological diversity among species is reflected by the number of and is a group of large-bodied and variably streaky spiderhunters subspecies (Dickinson ), from none (A. juliae, A. crassirostris, that occurs throughout Sundaland. Classification of the complex A. flavigaster) to  (A. longirostra). In morphology, spiderhunters are has been the subject of considerable debate, and opinions regard- larger and heavier-bodied than other sunbirds, and like female sun- ing the number, distribution, and rank of taxa vary substantially. birds they usually have drab greenish, yellow, or gray . One Here, we attempt to resolve the uncertainty in their relationships exception is Whitehead’s Spiderhunter (A. juliae), an endemic of Bor- and distribution by focusing primarily on the relationships of nean mountains that is dark brown with white streaks. All feed on A. affinis everetti and A. modesta modesta on Borneo. Sharpe () and small arthropods (not necessarily ). described A. everetti from the vicinity of Mt. Kinabalu (Sabah, OCTOBER 2011 — SPIDERHUNTER PHYLOGENY — 779

Malaysian north Borneo) and considered it an endemic species sequencing products were purified through Sephadex or precip- because of its large size and heavy streaking. Over the next cen- itation in % ethanol and then analyzed on an ABI  Prism tury, this taxon has variably been treated as a subspecies of an ex- xl Genetic Analyzer (Applied Biosystems). SEQUENCHER, panded A. affinis that also included all populations of A. modesta version . (Gene Codes, Ann Arbor, Michigan), was used to rec- (e.g., Chasen ; Smythies , ), as a Bornean endemic oncile complementary gene sequences and align sequences across species (Stresemann , Rand , Sibley and Monroe , taxa. MacKinnon and Phillipps , Inskipp et al. ), or as a sub- Ancient DNA was extracted in a clean lab dedicated to his- species of a more narrowly circumscribed A. affinis that excluded torical samples with strict protocols to prevent contamination. A. modesta (lowland Borneo, Sumatra, and mainland Asia) (Da- Historical samples were extracted with Qiagen kits using carrier vison , Dickinson , Mann ). The situation is further RNA and multiple negative controls. In addition, DNA was eluted complicated because A. modesta appears not to occur in Sabah twice with  μL of the provided buffer and both eluates were and A. affinis occupies both the lowlands and mountains in that combined and concentrated to ~ μL to increase DNA yield. part of Borneo (Chasen ; Sheldon et al. , a, b; Haines Ancient DNA was amplified and, if necessary, reamplified with ). Nevertheless, birdwatchers frequently report the two spe- PCR product. The PCR product was verified on agarose gels and, cies in sympatry in Sabah, and recent revisions continue to list if present, samples were cleaned with the Qiagen PCR Purifica- two species in north Borneo (Cheke and Mann , Mann ). tion kit, cycle sequenced, and cleaned with Sephadex prior to se- We compared DNA sequences of nuclear and mitochondrial quencing on an ABI  platform. Ancient DNA was sequenced genes from all described species of Arachnothera, including a at ND using the published primer sequences above as well as a broad geographic sample of the two widespread and variable spe- series of short, overlapping primer pairs that were designed using cies complexes (A. longirostra and A. affinis–modesta). Our main a degenerate consensus sequence obtained from contemporary goals were to evaluate the taxonomic status of several populations Arachnothera sequences. These short, degenerate internal primer and define species limits that accurately describe diversity in the pairs amplified consecutive – bp sections of the ND gene genus, and then apply patterns evident in the phylogeny to shed (see Table ) and were also used in various combinations to am- light on the history of diversification of the spiderhunters and plify larger fragments. Sequence coverage for historical samples other similarly distributed groups of the Greater Sunda and Phil- was approximately × and sequences within an individual were ippine islands.

TABLE 1. Sequences of newly designed primers for amplify- METHODS ing and sequencing ND2 from museum study-skin samples.

Samples of spiderhunters were collected by us and others across Primer name 5’ to 3’ sequence southeast and southern Asia (Appendix). Sampling was most dense in Borneo and the Philippines to assess populations in the ND2Int1F AACCCCMTAGCAAAASYA A. longirostra and A. affinis–modesta species complexes. The in- ND2Int1R TTRATTTCTAGKCCGGYT group consisted of  individuals representing all species in the ND2Int2F ACWACYATYACAGTAWCRAGCA ND2Int2R GRAAGTATTTRGTTGCGG genus. Most samples were derived from modern specimens col- ND2Int3F CCAYCCVCGAGCYATTGA lected in the wild. However, ancient DNA of A. juliae and two pop- ND2Int3R GRTGGGTTAGTTGGGTRA ulations in the A. affinis–modesta complex (A. affinis from Java ND2Int4F AGCATRACYAAYGCATGA and A. modesta from Sumatra) was extracted from museum study ND2Int4R TKACYGTGGATAGAAGRAG skins because no fresh samples were available. As outgroups, we ND2Int5F CTAGTMCCATTCCAYTTCT included a selection of other sunbird and flowerpecker (Dicaeidae) ND2Int5R TGGGGTTKARTGATKGKG genera. A leafbird (Chloropsis hardwickii) and a dipper (Cinclus ND2Int6F ATTCCCHCCWCTCACAYT pallasii) served as more distant outgroups to root trees. ND2Int6R GCKAGGATTTTTCGGATT Total genomic DNA from modern samples was extracted from frozen or alcohol-preserved muscle tissue using standard ND2Int7F TRGGMGGATGAATAGGAYT Qiagen extraction protocols (Qiagen, Valencia, California). Mo- ND2Int7R ATKAGYGAGTAYAGGTAGAAG lecular markers consisted of sequences of the mitochondrial ND2Int8F CTCMTACRACCCYAAACT genes nicotinamide adenine dinucleotide dehydrogenase sub- ND2Int8R TYGCRTTTADRGATGGRRYT units  (ND; , base pairs [bp]) and  (ND;  bp), and ND2Int9F AMCCACMYTAATAACWGCA the fifth nuclear intron of the transforming growth factor, β ND2Int9R TAGGGAGAGYAGGGTTAG (TGFβ;  bp aligned). These were amplified using the prim- ND2Int10F TACTAACCCTRCTCTCCC ers L–H (Hackett , Johnson and Sorenson ), ND2Int10R TTGYTTRGTYAGTTCYTGKAYG L–H (Chesser ), and TGF and TGF (Primmer et ND2Int11F ACAGGATTCMTKCCYAAA al. ), respectively. Polymerase chain reaction [PCR] products ND2Int11R GGCGKAGGTAGAAGAAYA were purified with ExoSAP-IT (US; Amersham Biosciences, ND2Int12F AATAKCCCTACTATCYCTAC Piscataway, New Jersey) or Qiagen Qiaquick PCR purification kits. ND2Int12R GRYWGGYTTRTTAGTRYG Purified PCR products were then sequenced with ABI Prism Big- ND2Int13F CCCYCCYCAYACMACAAA Dye Terminator chemistry, version . (Applied Biosystems, Fos- ter City, California), using the same primers used for PCR. Cycle ND2Int13R TGKRTADATKAKKGGDGCVRT 780 — MOYLE ET AL. — AUK, VOL. 128

verified to ensure that identical sequences were obtained for all base frequencies across taxa detected no significant differences combinations of published and internal primers. for each gene (ND, P = .; ND, P = .; TGFβ, P = .). Phylogenetic analysis was conducted using maximum like- Aligned TGFβ sequences contained several inferred in- lihood (ML) implemented in the software RAXML, version .. sertions or deletions (indels), but alignment of the sequences (Stamatakis et al. ). Sequences were divided into four parti- was fairly straightforward for two reasons. First, indels were tions (three mitochondrial codon positions and the intron) and infrequent enough that they generally did not overlap, allow- analyzed using the GTRGAMMA model of evolution. Clade sup- ing homologous indels to be identified. Second, the nucleotide port was evaluated with , fast bootstrap pseudoreplicates sequences themselves were not highly diverged, which simpli- (Stamatakis et al. ). We also conducted Bayesian analysis fied alignment and default placement of indels. Shared indels in- (BA) using Markov chain Monte Carlo (MCMC) tree searches cluded a -bp deletion in two individuals of both A. chrysogenys in MRBAYES, version .. (Huelsenbeck and Ronquist , and A. clarae; a -bp deletion in two individuals of A. longirostra Ronquist and Huelsenbeck ). MtDNA sequences were par- ( and ) and olivacea; a -bp insertion in the titioned by codon positions, and the nuclear intron was treated three individuals of Hypogramma; and a -bp deletion shared by as a separate partition. MRMODELTEST, version . (Nylander all individuals of A. longirostra, A. rosbusta, A. crassirostris, An- ), was used to determine the substitution model based on threptes singalensis, christinae, Nectarinia olivacea, Akaike’s information criterion. Two independent runs of  mil- N. sperata, and rectirostris. lion generations were conducted with default chain heating Phylogenetic results.—The taxon stability test revealed that conditions, sampling every , generations. Stationarity was samples with missing data were placed with the same, or more, evaluated with the program AWTY (Nylander et al. ) by consistency than samples with full data. The nine most unsta- plotting bipartition frequencies of each run in nonoverlapping ble taxa were the nine samples of Arachnothera everetti. These bins of , generations (“Slide”) and comparing the fre- samples had very similar sequences, and apparently changed quency of bipartitions between the two runs with various burn- relationships among bootstrap replicates, but were always re- in proportions removed (“Compare”). Generations before the covered as a clade. Bayesian and ML analysis produced differ- runs had stabilized and converged on similar posterior proba- ent topologies but they did not conflict at any well-supported bilities for clades were discarded as burn-in. The remaining trees nodes (>% boostrap support and . posterior probability). were retained and summarized in a majority-rule consensus However, analysis of each marker individually revealed conflict- tree. Bayesian analyses were also conducted on the mitochon- ing phylogenetic signal between the nuclear and mitochondrial drial and nuclear data separately to examine congruence in phy- data. The description of relationships that follows is based on the logenetic signal between the two markers. ML topology of the combined data (Fig. ). Differences in trees We examined the influence of missing data by plotting taxon between the markers are shown in Figure  and are described stability against the proportion of missing data for each individual following the discussion of general relationships of the - in MESQUITE (Maddison and Maddison ). This test has been hunters. Combined data analysis recovered the Purple-naped useful in identifying taxa that cannot be placed with confidence Sunbird (Hypogramma hypogrammicum) inside Arachnothera, because of missing data. For sample trees, we used bootstrap rep- rendering the genus paraphyletic. Support for the node that in- licates from the maximum likelihood analysis. cludes Hypogramma within Arachnothera was fairly strong, with a Bayesian posterior probability of . and ML boostrap- RESULTS ping support at %. This node unites Hypogramma with clade A of the spiderhunters (Fig. ). Clade A split into  subclades, Sequence attributes.—The DNA sequence matrix contained  one of which included the Bornean endemic A. juliae sister to ingroup individuals from  species and , bases (, ND, a clade consisting of A. chrysogenys and the Philippine endemic  ND,  TGFβ), of which  (.%) were parsimony infor- A. clarae. All relationships within this subclade received strong mative ( ND,  ND,  TGFβ). Sequences derived from support. Within the other A subclade, A. flavigaster branched museum study skins were limited to ND. All newly produced basally, followed by A. magna. TheA. affinis–modesta species sequences are available from the GenBank database (accession complex was strongly supported as sister to A. magna and con- numbers JF–JF). Aligned ND and ND sequences tained substantial geographic structure. It consisted of three lin- appeared to be of mitochondrial origin rather than nuclear copies, eages comprising individuals from Java, Sumatra–Sarawak, and they contained no stop codons, overlapping fragments contained the Bornean mountains and northeast Bornean lowlands. The no conflicts, base composition was relatively homogeneous across Javan individual was sister to the other two clades with strong taxa, codon positions contained expected relative divergences ML bootstrap support (%) and marginally significant poste- (>>), and resulting relationships contained no highly suspect rior probability (.). arrangements. Base composition in the mitochondrial data was The second clade of spiderhunters (Fig. ; clade B) contains biased, with an excess of adenine and cytosine (A = ., C = just three species. Strong nodal support united A. robusta and ., G = ., T = .), but consistent with compositional pat- A. crassirostris as sister taxa, with A. longirostra as sister to both of terns found in these genes in other bird groups (e.g., Sheldon et al. them. Arachnothera longirostra divided into three distinct clades , Moyle et al. ). Base composition in the intron data was corresponding to samples from Palawan, the southern Philip- more homogeneous (A = ., C = ., G = ., T = .). pines (Mindanao, Dinagat, and Bohol), and all non-Philippine Models of evolution chosen for all likelihood analyses accounted areas included in the study. The southern Philippine clade was for base composition bias. A chi-square test for homogeneity of separated by a deep branch from the Palawan population and its OCTOBER 2011 — SPIDERHUNTER PHYLOGENY — 781

FIG. 2. Maximum-likelihood estimate of the spiderhunter phylogeny based on combined mitochondrial and nuclear sequences. Branch support num- bers are bootstrap proportions/Bayesian posterior probability. Asterisks indicate 100% bootstrap and 1.0 Bayesian support. 782 — MOYLE ET AL. — AUK, VOL. 128

well-defined sister, the non-Philippine populations. Within the of the Nectariniidae. The nuclear data indicate that Arachnothera latter, individuals from Borneo, Malaya, and sorted into monophyly is further disrupted by the other sunbirds we included well-supported clades, although distinguished by shallow genetic in the study as outgoups. Intron comparisons place the five sun- divergences. bird species in a group that is sister to clade B of Arachnothera The main discrepancy between the nuclear and mitochon- with high support (Fig. ). The mtDNA data do not corroborate drial data concerned the placement of sunbirds other than Hy- this relationship, but instead place the five other sunbirds outside pogramma (i.e., Anthreptes singalensis, A. rectirostris, Nectarinia of Arachnothera–Hypogramma. The latter arrangement seems sperata, N. olivacea, and Aethopyga christinae). Mitochondrial most likely, given the many differences between Arachnothera– data placed them solidly outside of Arachnothera, as expected. By Hypogramma and other sunbirds (see below). The incongruence contrast, the nuclear intron data placed them sister to clade B of between mitochondrial and nuclear genes may be a case of deep Arachnothera with strong support. This relationship was further coalescence and gene tree–species tree discordance. More in- supported by a -bp deletion (Fig. ). The conflict in phylogenetic dependent markers will need to be applied to this problem to signals likely contributed to the low support for Arachnothera (in- resolve it. cluding Hypogramma) in the combined analysis. Support for Hy- A close relationship, or monophyly, of Hypogramma and pogramma within Arachnothera derived mostly from the mtDNA Arachnothera is consistent with their distribution, morphology, data, which placed it clearly as sister to clade A (Fig. ). The nuclear and behavior. Hypogramma occurs in the same areas as Arach- data placed Hypogramma in a polytomy with four other clades nothera (i.e., throughout much of Southeast Asia and the Greater (Fig. ). Sundas, except Java). Both sexes of Hypogramma and several spe- Uncorrected proportional genetic distances (p distances) be- cies of spiderhunters have varying degrees of underpart streaking. tween samples can be used as a rough heuristic for divergence be- Such streaking occurs only in females of a few other sunbird spe- tween lineages. For the ND data, A. longirostra individuals from cies. Hypogramma’s only overt plumage difference from spider- Borneo were .% diverged from the peninsular Malaysian indi- hunters is the small iridescent purple patch on the nape and rump vidual, .% diverged from the Vietnam birds, .–.% from the of the male. Hypogramma constructs a purse-shaped nest typical Palawan birds, and about % from the southern Philippine clade. of sunbirds, but it is suspended from the underside of a large leaf as Individuals of A. affinis–modesta from the Bornean mountains in spiderhunters (Wells , Cheke and Mann ). In spider- and northeast lowlands were .% diverged from Sumatran and hunters, both parents incubate eggs, which is unusual in sunbirds. Sarawak lowland birds, and .% diverged from the Javan bird. Hypogramma’s incubation behavior is unknown. Hypogramma is The Sumatran individual was only .–.% diverged from mul- smaller than several species of spiderhunters but is roughly the tiple individuals from the Sarawak lowlands. The maximum diver- same size as A. longirostra and A. crassirostris. If this phylogenetic gence across five individuals of A. magna was almost %, between result is corroborated, we suggest that Hypogramma be subsumed a bird from Vietnam () and samples from , Myanmar, within Arachnothera. and Vietnam. Curiously, another individual () collected Higher-level relationships of spiderhunters are well resolved, from the same site on the same trip to Vietnam (P. Sweet pers. but they do not indicate clear geographic or diversification pat- comm.) grouped with the samples from Myanmar and China. The terns, except in the case of A. affinis, A. modesta,and A. magna. cause of such a high divergence within this Vietnam population There are three groups of relatively closely related spiderhunter is unknown. species: () A. affinis, A. modesta, A. magna, and A. flavigaster (Fig. ; clade A); () A. clarae, A. chrysogenys, and A. juliae (Fig. ; clade DISCUSSION A); and () A. crassirostris, A. robusta, and A. longirostra (Fig. ; clade B). The first consists of three medium-sized and one large- Higher-level phylogeny.—Analysis of the nuclear and mitochon- sized species. The three medium-sized species—A. affinis, A. mod- drial sequence data produced conflicting hypotheses about the esta,and A. magna—are all streaked, and they replace one another relationships of Arachnothera. Although this conflict precluded geographically. Where they co-occur, A. modesta inhabits the reliable inference of some of the deeper relationships among ma- lowlands and A. magna (Malay Peninsula) or A. affinis (Borneo) jor spiderhunter clades, most nodes among species received strong occupies the highlands (but see below for more details on the tax- support from both markers, allowing us to discuss several aspects onomy and distribution of A. affinis). This pattern suggests a spe- of spiderhunter and hypotheses of avian diversification cies group that subdivided and then came into secondary contact. in the Sunda region. The second group consists of three medium-sized species. One of Spiderhunters do not appear to be monophyletic. The them is widespread in Sundaland (A. chrysogenys) and two are dis- Purple-naped Sunbird is nested inside Arachnothera and is the tributionally restricted endemics, A. clarae in the Philippines and sister of one of the two main spiderhunter clades. Support for A. juliae in the mountains of Borneo. This distribution is highly the novel phylogenetic placement of Hypogramma derives mostly reminiscent of a clade of trogons in the genus Harpactes (Hos- from the mtDNA data; the intron places it in a polytomy with sev- ner et al. ), described below. The third group contains the two eral other lineages (Fig. ). This result is unexpected, given that smallest spiderhunters, A. longirostra and A. crassirostris, as well we can find no mention of possible inclusion of Hypogramma as the somewhat larger but morphologically similar A. robusta. within spiderhunters in the literature. In the only other molec- Taxonomy of species groups.—Both genetic markers indicated ular phylogenetic study that included both genera, Sibley and deep divergences within A. longirostra that distinguished the two Ahlquist’s () DNA–DNA hybridization comparisons recov- endemic Philippine taxa from all other Little Spiderhunter popula- ered Arachnothera and Hypogramma in a polytomy with the rest tions included in the study. The two Philippine taxa are the Palawan OCTOBER 2011 — SPIDERHUNTER PHYLOGENY — 783 esian posterior probabilities. Asterisks indicate Asterisks posterior indicate probabilities. esian . support and (B) Branch trees DNA numbers of (A) nuclear inference. Comparison are Bay constructed mitochondrial 3. Bayesian by IG 1.0 Bayesian support. F 784 — MOYLE ET AL. — AUK, VOL. 128

subspecies dilutior and the southern Philippine subspecies flam- A. a. affinis of Java and A. a. everetti of montane and northeast low- mifera (and presumably randi from Basilan Island, which we did land Borneo, is not a natural group and should not be maintained. not sample). Their genetic distinctiveness was suggested earlier by Although we predicted that these two taxa would be closely related Rahman et al. () and Lohman et al. (), and for this reason because other montane Bornean bird species appear to have close we investigated them with expanded sampling. In addition to their Javan relatives (Sheldon et al. a), this prediction proved to be molecular divergence, dilutior and flammifera are also morphologi- wrong. Outside of Borneo, A. a. affinis and the widespread lowland cally distinct (Table ). Subspecies dilutior has a yellow eye-ring that A. modesta have at times been considered conspecific (Chasen , fades in museum specimens and is not depicted in field guides, al- Rand ), but lumping those taxa would also create a paraphy- though it is noted in the text by Kennedy et al. (). It is drabber letic group. Instead, our comparisons defined three lineages, each than populations in neighboring Borneo, having almost no yellow of which should be considered distinct species. Arachnothera affinis in its underparts. Subspecies flammifera is small-bodied, and its should be split into A. affinis from Java and Bali and A. everetti from yellow plumage is restricted to the lower abdomen, flanks, and vent. montane and northeast lowland Borneo. The constitution of A. mod- Both of the Philippine taxa lack light lores and eyebrows, typical of esta (Davison ) remains largely unchanged. The lowland individ- A. longirostra in Sundaland and continental Southeast Asia. As a re- uals that we compared from Sarawak in western Borneo were most sult of molecular and morphological differences, the endemic Phil- similar to those from Sumatra, which suggests a strong Sundaic con- ippine little spiderhunters should be treated as two distinct species: nection (Lim et al. ). Finally, there is no evidence that A. modesta A. dilutior Sharpe, , on Palawan; and A. flammifera Tweeddale, occurs in Sabah. The seven individual streaky spiderhunters that we , on the southern Philippine islands. Within Borneo, no break compared from Sabah, although derived from sites spanning – appears between the birds of Sarawak and Sabah, as occurs in many , m in elevation, were all closely related to an individual from other species (Moyle et al. , Lim et al. ). Panmixia may be montane Sarawak rather than individuals from the lowlands of Sar- attributed to the high vagility of spiderhunters, which search widely awak or Sumatra. Future records from Sabah should be considered for food and live in a variety of habitats. A. everetti unless evidence of A. modesta is thoroughly documented. Current taxonomy inaccurately reflects relationships within Diversification in the Sunda region.—Species of spiderhunters the A. affinis species complex. Arachnothera affinis, which includes differ substantially at mitochondrial loci, and several have largely overlapping distributions (Fig. ). These factors suggest that the forces behind the diversification of major groups of spiderhunters TABLE 2. Proposed species-level taxa of spiderhunters (genus Arachnothera) (and perhaps Hypogramma) are ancient and not the result of recent with descriptions and distributional information. geological and environmental events such as the eustatic sea-level Latin name Description a Distribution changes caused by Pleistocene glaciation. The widespread sympatry of most spiderhunter species in Sundaland (Fig. ) suggests that, A. dilutior Yellow eye-ring, drabber under- Palawan once diverged, these highly vagile birds took advantage of recent parts than A. longirostra, no land bridges to move back and forth among islands (unfortunately, light eyebrow or lores as in we did not have samples to examine genetic distances among popu- A. longirostra A. flammifera Small size, yellow restricted to Southern Philippines lations on different Sunda islands). Another interesting aspect of spi- lower abdomen, flanks, and (Bohol, Dinagat, Leyte, derhunter distribution is that Sumatra lacks a montane specialist; vent; lacks light eyebrows and Mindanao, Samar, and this suggests that the montane taxa (A. everetti and A. magna) did lores of A. longirostra probably Basilan) not make it that far south or west in Sundaland or went extinct there. A. everetti Large-bodied, with dark gray- Montane Borneo and With respect to distribution and genetic divergence, the spi- brown streaking from chin to northeast lowland derhunters strongly resemble the trogons (Harpactes), another abdomen and flanks; streaks Borneo Asian group for which we have molecular phylogenetic data. more diffuse on abdomen; Harpactes species have much the same distribution and their ma- dull yellow wash to abdomen, jor clades are separated by long stem lineages that lead to several marked sexual size dimor- phism; wing (male) 87–96 mm broadly sympatric species (Hosner et al. ). One trogon pattern (mean 91.2, n = 14), (female) is particularly reminiscent of the spiderhunters: a clade of three 78–85 mm (mean 81.4, n = 9) taxa consisting of one broadly distributed species (A. chrysogenys; A. affinis Similar to A. everetti, but with Java and Bali H. kasumba), one Bornean mountain endemic (A. juliae; H. white- rusty breast streaks; wing headi), and one Philippine endemic (A. clarae; H. ardens). (male) 76–94 mm (mean 84.9, At the population level, we are beginning to see several vari- n = 8), (female) 80–85 mm ations on a common theme in Sundaland, in which populations (mean 82.5, n = 4) in the lowlands of northeast Borneo are genetically (and some- A. modesta Much smaller body, shorter bill, Extreme southern times morphologically) distinct from those in the lowlands else- and more lightly streaked Myanmar and where in Borneo (Moyle et al. ; Sheldon et al. b; Lim et al. than A. everetti; wing (male) Thailand, Peninsular 77–82 mm (79.5, n = 2) Malaysia, Sumatra, , ). These observations add to a growing body of evidence lowland Borneo except of a major biogeographic break within the island corresponding in the northeast roughly with the border of Sabah (Gorog et al. , Ryan and Esa ). The northern portion of Borneo’s central highlands aMeasurements of A. everetti and A. affinis are from Davison (1999). Measurement of A. modesta is of two Louisiana State University specimens from Mt. Pueh, western runs along the border between Sabah and Sarawak. This mon- Sarawak. tane barrier, along with the isolating effects of a number of rivers OCTOBER 2011 — SPIDERHUNTER PHYLOGENY — 785

that flow from the northern face of the central highland into the A savanna corridor in Sundaland? Quaternary Science Reviews South China sea, could present a formidable dispersal impedi- :–. ment to lowland taxa. Interestingly, even when Sundaland was Campbell, P., C. J. Schneider, A. M. Adnan, A. Zubaid, maximally exposed during glacial periods, little additional land and T. H. Kunz. . Comparative population structure connected Sabah to the rest of Borneo because of the deep wa- of Cynopterus bats in peninsular Malaysia and southern ters around Sabah. Such insularity could further promote the dif- Thailand. Molecular Ecology :–. ferentiation of Sabah populations. Moyle et al. () discovered Cannon, C. H., R. J. Morley, and A. B. G. Bush. . The current the pattern of two lowland populations in the White-crowned refugial rainforests of Sundaland are unrepresentative of their bio- Forktail but also found a third, more distant taxon (E. borneensis) geographic past and highly vulnerable to disturbance. Proceedings in the mountains of Borneo. Moreover, the montane E. borneensis of the National Academy of Sciences USA :–. did not exhibit a biogeographic break corresponding to the low- Chasen, F. N. . A handlist of Malaysian birds. Bulletin of the land break. In the present study, we found a variation on this pat- Raffles Museum :–. tern. As in the forktails, the montane A. everetti is distinct from Cheke, R. A., and C. F. Mann. . Family Nectariniidae (sun- the lowland A. modesta, but there is no lowland population of A. birds). Pages – in Handbook of the Birds of the World, vol. modesta in Sabah; it is replaced there by A. everetti, which inhabits : Penduline-tits to Shrikes (J. del Hoyo, A. Elliott, and D. Chris- both montane and lowland areas. Whether the different genetic tie, Eds.). Lynx Edicions, Barcelona, Spain. patterns among these species groups (three vs. two distinct Bor- Chesser, R. T. . Molecular systematics of the rhinocryptid nean lineages) indicate two distinct historical causes of diversi- genus Pteroptochos. Condor :–. fication, or rather variation in the outcome of a single historical Davison, G. W. H. . Notes on the taxonomy of some Bornean driver, is uncertain. Quite possibly, A. everetti was isolated in a birds. Sarawak Museum Journal :–. northern Bornean rainforest refuge, as apparently were other low- Diamond, J. M., and M. E. Gilpin. . Biogeographic umbilici land Sabahan birds (Lim et al. , ). It maintained that low- and the origin of the Philippine avifauna. Oikos :–. land distribution in the absence of A. modesta and also expanded Dickinson, E. C. . The Howard and Moore Complete Check- southward through the mountains, where A. modesta was absent. list of the Birds of the World, rd ed. Christopher Helm, London. Alternatively, A. everetti was always a montane species that was Esselstyn, J. A., and C. H. Oliveros. . Colonization of the able to expand into the lowlands of Sabah because of the absence Philippines from Taiwan: A multi-locus test of the biogeographic of A. modesta, but not into the lowlands elsewhere in Borneo and phylogenetic relationships of isolated populations of shrews. where A. modesta occurred. It is also possible that A. everetti once Journal of Biogeography :–. inhabited a broad elevational range across all of Borneo but has Esselstyn, J. A., C. H. Oliveros, R. G. Moyle, A. T. Peterson, recently been displaced from the western lowlands by A. modesta. J. A. McGuire, and R. M. Brown. . Integrating phyloge- netic and taxonomic evidence illuminates complex biogeographic ACKNOWLEDGMENTS patterns along Huxley’s modification of Wallace’s Line. Journal of Biogeography :–. The following museums kindly provided tissue samples: Univer- Flenley, J. R. . Tropical forests under the climates of the last sity of Kansas Natural History Museum, Louisiana State University , years. Climatic Change :–. Museum of Natural Science, American Museum of Natural History Gorog, A. J., M. H. Sinaga, and M. D. Engstrom. . Vicariance (AMNH), and Cincinnati Museum Center. We particularly thank or dispersal? Historical biogeography of three Sunda shelf murine the AMNH, the Academy of Natural Sciences of Philadelphia, the rodents (Maxomys surifer, Leopoldamys sabanus and Maxomys Yale Peabody Museum, and the Western Foundation of Vertebrate whiteheadi). Biological Journal of the Linnean Society :–. Zoology for providing toe pad samples of inaccessible populations. Hackett, S. J. . Molecular phylogenetics and biogeography of For permission to undertake research in the Philippines, we thank tanagers in the genus Ramphocelus (Aves). Molecular Phylogenet- the Department of Environment and Natural Resources of the Phil- ics and Evolution :–. ippines and the Within Protected Areas and Wildlife Bureau. For Haines, C. L. . Comparative phylogeography of four montane research permission and assistance in Malaysia, we thank the Prime bird species in Sabah, Malaysian Borneo. M.S. thesis, Louisiana Minister’s Department, the Chief Minister’s Department of Sabah State University, Baton Rouge. and Sarawak, the Sabah Wildlife Department, Sabah Parks, Sabah Hall, R., and J. D. Holloway. . Biogeography and Geological Museum, Sarawak Department, Sarawak Forestry Corpora- Evolution of SE Asia. Backhuys, Leiden, The Netherlands. tion, Sarawak Biodiversity Institute, University Malaysia Sarawak, Heaney, L. R. . Biogeography of mammals in SE Asia: Esti- and Grand Perfect Sdn. Bhd. For assistance in the lab, we thank mates of rates of colonization, and . Biologi- B. Cerame and M. Bowen. The research was funded by the Coypu cal Journal of the Linnean Society :–. Foundation, National Geographic Society, and the National Sci- Heaney, L. R. . A synopsis of climatic and vegetational change ence Foundation (DEB- to F.H.S., DEB- and DEB- in Southeast Asia. Climatic Change :–.  to R.G.M.). Hosner, P. A., F. H. Sheldon, H. C. Lim, and R. G. Moyle. . Phylogeny and biogeography of the Asian trogons (Aves: Tro- LITERATURE CITED goniformes) inferred from nuclear and mitochondrial DNA sequences. Molecular Phylogenetics and Evolution :–. Bird, M. I., D. Taylor, and C. Hunt. . Palaeoenviron- Huelsenbeck, J. P., and F. Ronquist. . MRBAYES: Bayesian ments of insular Southeast Asia during the last glacial period: inference of phylogenetic trees. Bioinformatics :–. 786 — MOYLE ET AL. — AUK, VOL. 128

Inskipp, T., N. Lindsey, and W. Duckworth. . An Anno- Primmer, C. R., T. Borge, J. Lindell, and G.-P. Sætre. . tated Checklist of the Birds of the Oriental Region. Oriental Bird Single-nucleotide polymorphism characterization in species with Club, Bedfordshire, United Kingdom. limited available sequence information: High nucleotide diversity Jansa, S. A., F. K. Barker, and L. R. Heaney. . The pattern revealed in the avian genome. Molecular Ecology :–. and timing of diversification of Philippine endemic rodents: Evi- Rahman, M. A., D. F. A. Gawin, and C. Moritz. . Patterns of dence from mitochondrial and nuclear gene sequences. System- genetic variation of the Little Spiderhunter (Arachnothera longirostra) atic Biology :–. in Southeast Asia. Raffles Bulletin of Zoology :–. Johnson, K. P., and M. D. Sorenson. . Comparing molec- Rand, A. L. . Family Nectariniidae. Pages – in Check- ular evolution in two mitochondrial protein coding genes of the World, vol.  (R. A. Paynter, Ed.). Museum of (cytochrome b and ND) in the dabbling ducks (Tribe: Anatini). Comparative Zoology, Cambridge, Massachusetts. Molecular Phylogenetics and Evolution :–. Ronquist, F., and J. P. Huelsenbeck. . MrBayes : Bayes- Kennedy, R. S., P. C. Gonzales, E. C. Dickinson, H. C. ian phylogenetic inference under mixed models. Bioinformatics Miranda, and T. H. Fisher. . A Guide to the Birds of the :–. Philippines. Oxford University Press, New York. Ryan, J. R. J., and Y. B. Esa. . Phylogenetic analysis of Ham- Lim, H. C., M. A. Rahman, S. L. H. Lim, R. G. Moyle, and F. H. pala fishes (Subfamily Cyprininae) in Malaysia inferred from Sheldon. . Revisiting Wallace’s haunt: Coalescent simula- partial mitochondrial cytochrome b DNA sequences. Zoological tions and comparative niche modeling reveal historical mecha- Science :–. nisms that promoted avian population divergence in the Malay Sangster, G., P. Alström, E. Forsmark, and U. Olsson. . Archipelago. Evolution :–. Multi-locus phylogenetic analysis of chats and fly- Lim, H. C., F. Zou, S. S. Taylor, B. D. Marks, R. G. Moyle, G. catchers reveals extensive paraphyly at family, subfamily and Voelker, and F. H. Sheldon. . Phylogeny of magpie- genus level (Aves: Muscicapidae). Molecular Phylogenetics and robins and shamas (Aves: Turdidae: Copsychus and Trichixos): Evolution :–. Implications for island biogeography in Southeast Asia. Journal Sharpe, R. B. . Bornean notes. I. First list of birds from Mt. of Biogeography :–. Kalulong, in Sarawak. II. A list of the birds collected by Mr. A. H. Lohman, D. J., K. K. Ingram, D. M. Prawiradilaga, K. Winker, Everett on Mt. Penrisen and Mt. Poeh, in Sarawak. III. Descrip- F. H. Sheldon, R. G. Moyle, P. K. L. Ng, P. S. Ong, L. K. Wang, tion of a new Spilornis from Borneo. IV. A note on the Baza of T. M. Braile, and others. . Cryptic genetic diversity in Borneo. V. Notes on Mr. A. H. Everett’s collections of birds from “widespread” Southeast Asian bird species suggests that Philip- northern Borneo and Sarawak. VI. Additions to the avifauna of pine avian endemism is gravely underestimated. Biological Con- Mount Kina Balu. VII. Description of the nest and eggs of Staph- servation :–. idia everetti. Ibis :–. MacKinnon, J., and K. Phillipps. . A Field Guide to the Sheldon, F. H., H. C. Lim, J. Nais, M. Lakim, A. Tuuga, P. Birds of Borneo, Sumatra, Java, and Bali. Oxford University Press, Malim, J. Majuakim, A. Lo, M. Schilthuizen, P. A. Hos- Oxford, United Kingdom. ner, and R. G. Moyle. a. Observations on the ecology, Maddison, W. P., and D. R. Maddison. . MESQUITE: A distribution, and biogeography of forest birds in Sabah, Malay- modular system for evolutionary analysis. [Online.] Available at sia. Raffles Bulletin of Zoology :–. mesquiteproject.org. Sheldon, F. H., D. J. Lohman, H. C. Lim, F. Zou, S. M. Good- Mann, C. F. . The Birds of Borneo, an Annotated Checklist. man, D. M. Prawiradilaga, K. Winker, T. M. Braile, and British Ornithologists’ Union and British Ornithologists’ Club, R. G. Moyle. b. Phylogeography of the magpie–robin spe- Peterborough, United Kingdom. cies complex (Aves: Turdidae: Copsychus) reveals a Philippine Meijaard, E. . Mammals of south-east Asian islands and their Late species, an interesting isolating barrier and unusual dispersal Pleistocene environments. Journal of Biogeography :–. patterns in the Indian Ocean and Southeast Asia. Journal of Bio- Moyle, R. G., J. Fuchs, E. Pasquet, and B. D. Marks. . Feed- geography :–. ing behavior, toe count, and the phylogenetic relationships among Sheldon, F. H., R. G. Moyle, and J. Kennard. . Ornithology alcedinine kingfishers (Alcedininae). Journal of Avian Biology of Sabah: History, gazetteer, annotated checklist, and bibliogra- :–. phy. Ornithological Monographs, no. . Moyle, R. G., M. Schilthuizen, M. A. Rahman, and F. H. Sheldon, F. H., L. A. Whittingham, R. G. Moyle, B. Slikas, Sheldon. . Molecular phylogenetic analysis of the White- and D. W. Winkler. . Phylogeny of swallows (Aves: crowned Forktail Enicurus leschenaulti in Borneo. Journal of Hirundinidae) estimated from nuclear and mitochondrial DNA Avian Biology :–. sequences. Molecular Phylogenetics and Evolution :–. Nylander, J. A. A. . MrModeltest, version . Program dis- Sibley, C. G., and J. E. Ahlquist. . Phylogeny and Classifi- tributed by the author. Evolutionary Biology Centre, Uppsala cation of Birds: A Study in Molecular Evolution. Yale University University, Uppsala, Sweden. [Online.] Available at www.abc. Press, New Haven, Connecticut. se/~nylander/mrmodeltest/mrmodeltest.html. Sibley, C. G., and B. L. Monroe, Jr. . Distribution and Taxon- Nylander, J. A. A., J. C. Wilgenbusch, D. L. Warren, and D. L. omy of Birds of the World. Yale University, New Haven, Connecticut. Swofford. . AWTY (are we there yet?): A system for graph- Smythies, B. E. . An annotated checklist of the birds of Borneo. ical exploration of MCMC convergence in Bayesian phylogenet- Sarawak Museum Journal :–. ics. Bioinformatics :–. Smythies, B. E. . The Birds of Borneo. Oliver and Boyd, London. Oliveros, C. H., and R. G. Moyle. . Origin and diversifica- Stamatakis, A., P. Hoover, and J. Rougemont. . A fast tion of Philippine bulbuls. Molecular Phylogenetics and Evolu- bootstrapping algorithm for the RAxML web-servers. Systematic tion :–. Biology :–. OCTOBER 2011 — SPIDERHUNTER PHYLOGENY — 787

Stresemann, E. . Vögel vomm Fluss Kayan (Nordost-Borneo). Whitmore, T. C., Ed. . Biogeographic Evolution of the Malay Temminckia :–. Archipelago. Clarendon Press, Oxford. Wells, D. R. . The Birds of the Thai-Malay Peninsula, vol. : . Christopher Helm, London. Associate Editor: J. P. Dumbacher

APPENDIX. Samples included in the study. Sources: Louisiana State University Museum of Natural Science (LSUMNS), University of Kansas Natural His- tory Museum (KUNHM), American Museum of Natural History (AMNH), Cincinnati Museum Center (CMC), Academy of Natural Sciences Philadelphia (ANSP), and Western Foundation of Vertebrate Zoology (WFVZ). Asterisks indicate samples extracted from museum study skins, for which only ND2 was sequenced.

Species Locality Elevation Sample number Collector–preparator

Arachnothera longirostra Crocker Range HQ, Sabah, Malaysia 1,000 m LSUMNS B36306 R. G. Moyle Serinsim, Sabah, Malaysia 200 m LSMMNS B46985 R. G. Moyle Bukit Hanta FR, Johor, Malaysia 80 m LSUMNS B52603 H. C. Lim Puerto Princesa, Palawan, Philippines 200 m KUNHM 12621 C. H. Oliveros Dinagat Island, Philippines 200 m KUNHM 14039 R. Fernandez Zamboanga, Mindanao, Philippines 80 m KUNHM 18123 C. H. Oliveros Puerto Princesa, Palawan, Philippines 200 m KUNHM 12706 C. H. Oliveros Mantalingajan Range, Palawan, Philippines 850 m KUNHM 12773 R. G. Moyle Dinagat Island, Philippines 200 m KUNHM 14174 R. Fernandez Bohol Island, Philippines 350 m KUNHM 20950 C. H. Oliveros Mt. Magdiwata, Mindanao, Philippines 500 m KUNHM 19010 C. H. Oliveros Penampang, Sabah, Malaysia 375 m KUNHM 17706 R. G. Moyle Ulu Kimanis, Sabah, Malaysia 550 m KUNHM 17768 R. G. Moyle Bintulu, Sarawak, Malaysia 150 m KUNHM 12343 R. G. Moyle Quang Nam, Vietnam 200 m AMNH DOT 10813 P. R. Sweet Quang Nam, Vietnam 200 m AMNH DOT 12304 R. T. Chesser A. crassirostris Kg. Pueh, Sarawak, Malaysia 60 m KUNHM 24436 R. G. Moyle A. robusta Mt. Trus Madi, Sabah, Malaysia 1,400 m LSUMNS B36483 R. G. Moyle Mt. Lucia, Sabah, Malaysia 900 m LSUMNS B51150 B. D. Marks A. flavigaster Ulu Kimanis, Sabah, Malaysia 550 m LSUMNS B61604 F. H. Sheldon Ulu Kimanis, Sabah, Malaysia 550 m KUNHM 17772 R. G. Moyle A. chrysogenys Serinsim, Sabah, Malaysia 180 m LSUMNS B47054 R. G. Moyle Ulu Kimanis, Sabah, Malaysia 550 m LSUMNS B61603 F. H. Sheldon Ulu Kimanis, Sabah, Malaysia 550 m KUNHM 17772 R. G. Moyle A. clarae Mindanao, Philippines CMC B2049 R. S. Kennedy Mindanao, Philippines CMC B36996 R. S. Kennedy Aurora, Luzon, Philippines 525 m KUNHM 19622 C. H. Oliveros A. modesta Kuching NP, Sarawak, Malaysia Sea level LSUMNS B52174 H. C. Lim Kg. Pueh, Sarawak, Malaysia 50 m LSUMNS B58549 F. H. Sheldon Kg. Pueh, Sarawak, Malaysia 50 m LSUMNS B58584 F. H. Sheldon Kg. Pueh, Sarawak, Malaysia 50 m LSUMNS B58637 F. H. Sheldon Meloewak, Sumatra ANSP 140139* Vanderbilt Expedition A. affinis Palabaeau Ratoe, West Java ANSP 56776* A. everetti Tawau Hills, Sabah, Malaysia 300 m LSUMNS B36402 R. G. Moyle Crocker Range HQ, Sabah, Malaysia 1,000 m LSUMNS B36310 R. G. Moyle Serinsim, Sabah, Malaysia 180 m LSUMNS B47062 F. H. Sheldon Mt. Lucia, Sabah, Malaysia 1,000 m LSUMNS B51142 R. G. Moyle Mt. Kinabalu, Sabah, Malaysia 2,100 m KUNHM 17801 R. G. Moyle Bario, Kelabit Plateau, Sarawak, Malaysia 1,690 m AMNH 648559* T. Harrisson Ulu Kimanis, Sabah, Malaysia 550 m KUNHM 17761 R. G. Moyle Temburong, LSUMZ 155705* A. Chapman Crocker Range HQ, Sabah, Malaysia 1,000 m LSUMNS B36309 R. G. Moyle A. magna Guanxi, China 500 m KUNHM 10401 T. J. Davis Guanxi, China 978 m KUNHM 10194 A. Nyari Kyi Tan, Myanmar KUNHM 15259 F. Steinheimer Quang Nam, Vietnam 1,450 m AMNH DOT 12277 R. T. Chesser Quang Nam, Vietnam 1,450 m AMNH DOT 12295 R. T. Chesser A. juliae Moyog, Sabah, Malaysia 1,470 m WFVZ 41711* D. Foote

(continued) 788 — MOYLE ET AL. — AUK, VOL. 128

APPENDIX. Continued.

Species Locality Elevation Sample number Collector–preparator Outgroups Anthreptes singalensis Klias Forest Reserve, Sabah, Malaysia 20 m LSUMNS B47156 B. D. Marks A. rectirostris Equatorial Guinea KUNHM 8499 L. A. Sánchez-González Aethopyga christinae Guanxi, China 500 m KUNHM 10276 M. B. Robbins Nectarinia sperata Ulu Tungud FR, Sabah, Malaysia 100 m LSUMNS B57444 R. G. Moyle N. olivacea Cameroon 650 m LSUMNS B27073 D. Dittmann Hypogramma hypogrammicum Tawau Hills, Sabah, Malaysia 150 m LSUMNS B36404 R. G. Moyle H. hypogrammicum Tawau Hills, Sabah, Malaysia 150 m LSUMNS B36422 R. G. Moyle H. hypogrammicum Tawau Hills, Sabah, Malaysia 150 m LSUMNS B38549 R. G. Moyle Dicaeum monticolum Meligan Range, Sabah, Malaysia 1,600 m KUNHM 17745 R. G. Moyle Prionochilus maculatus Bintulu, Sarawak, Malaysia 150 m KUNHM 12405 R. G. Moyle Chloropsis hardwickei Guanxi, China 978 m KUNHM 10019 A. T. Peterson–A. Nyari Cinclus pallasii China KUNHM 11241 B. W. Benz